Show simple item record

dc.contributor.authorIglesias, Gonzalo
dc.contributor.authorTambellini, William
dc.contributor.authorde Gispert, Adrià
dc.contributor.authorHasler, Eva
dc.contributor.authorByrne, William
dc.date.accessioned2019-01-15T00:30:25Z
dc.date.available2019-01-15T00:30:25Z
dc.identifier.urihttps://www.repository.cam.ac.uk/handle/1810/287962
dc.description.abstractWe describe a batched beam decoding algorithm for NMT with LMBR n-gram posteriors, showing that LMBR techniques still yield gains on top of the best recently reported results with Transformers. We also discuss acceleration strategies for deployment, and the effect of the beam size and batching on memory and speed.
dc.titleAccelerating NMT Batched Beam Decoding with LMBR Posteriors for Deployment
dc.typeConference Object
prism.publicationNamehttp://aclweb.org/anthology/N18-1000
dc.identifier.doi10.17863/CAM.35282
dcterms.dateAccepted2018-03-28
rioxxterms.versionofrecord10.17863/CAM.35282
rioxxterms.versionAM
rioxxterms.licenseref.urihttp://www.rioxx.net/licenses/all-rights-reserved
rioxxterms.licenseref.startdate2018-03-28
rioxxterms.typeConference Paper/Proceeding/Abstract
pubs.conference-nameProceedings of the North Americal Association of Computational Linguistics and Human Language Technologies Conference (NAACL-HLT ) 2018
pubs.conference-start-date2018-06-01
cam.orpheus.counter58
pubs.conference-finish-date2018-06-06
rioxxterms.freetoread.startdate2022-01-14


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record